Game Semantics: Easy as Pi

TL;DR

This paper introduces a new methodology called programming game semantics that simplifies game semantics by translating programming languages into a session typed pi-calculus, making complex models more accessible and easier to reason about.

Contribution

It presents a novel syntactic translation approach into a session typed pi-calculus, enabling easier reasoning and the first causal, non-angelic game model for a concurrent language.

Findings

First causal, non-angelic game model of CML

Adequate translation of CML into PiDiLL

Prototype implementation exploring OCaml subset

Abstract

Game semantics has proven to be a robust method to give compositional semantics for a variety of higher-order programming languages. However, due to the complexity of most game models, game semantics has remained unapproachable for non-experts. In this paper, we aim at making game semantics more accessible by viewing it as a syntactic translation into a session typed pi-calculus, referred to as metalanguage, followed by a semantics interpretation of the metalanguage into a particular game model. The syntactic translation can be defined for a wide range of programming languages without knowledge of the particular game model used. Simple reasoning on the model (soundness, and adequacy) can be done at the level of the metalanguage, escaping tedious technical proofs usually found in game semantics. We call this methodology programming game semantics. We design a metalanguage (PiDiLL) inspired from Differential Linear Logic (DiLL), which is concise but expressive enough to support features required by concurrent game semantics. We then demonstrate our methodology by yielding the first causal, non-angelic and interactive game model of CML, a higher-order call-by-value language with shared memory concurrency. We translate CML into PiDiLL and show that the translation is adequate. We give a causal and non-angelic game semantics model using event structures, which supports a simple semantics interpretation of PiDiLL. Combining both of these results, we obtain the first interactive model of a concurrent language of this expressivity which is adequate with respect to the standard weak bisimulation, and fully abstract for the contextual equivalence on second-order terms. We have implemented a prototype which can explore the generated causal object from a subset of OCaml.